RESEARCH Open Access

Glioma stem cells-derived exosomal miR-26a promotes angiogenesis of microvesselendothelial cells in gliomaZhi-Fei Wang*, Fan Liao, Hao Wu and Jin Dai

Abstract

Background: Cancer stem cells (CSCs), which are involved in cancer initiation and metastasis, could potentially releaseexosomes that mediate cellular communication by delivering microRNAs (miRNAs). Based on the role of miR-26a inangiogenesis of glioma, our study was performed to investigate whether glioma stem cells (GSCs)-derived exosomescontaining miR-26a could exert effects on angiogenesis of microvessel endothelial cells in glioma, in order to provide anew therapeutic RNA vehicle for glioma therapies.

Methods: The expression of miR-26a and PTEN in glioma was quantified and the interaction among miR-26a, PTEN andPI3K/Akt signaling pathway was examined. Next, a series of gain- and loss-of function experiments were conducted todetermine the role of miR-26a in angiogenesis of human brain microvascular endothelial cells (HBMECs). Subsequently,HBMECs were exposed to exosomes derived from GSCs with the gain−/loss-of-function of miR-26a. Finally, the effect ofexosomal miR-26a on angiogenesis of HBMECs was assessed both in vitro and in vivo.

Results: The results revealed that PTEN was down-regulated, while miR-26a was up-regulated in glioma. miR-26aactivated the PI3K/Akt signaling pathway by targeting PTEN. Restored miR-26a promoted proliferation, migration, tubeformation, and angiogenesis of HBMECs in vitro. In addition, GSCs-derived exosomes overexpressing miR-26a contributedto enhanced proliferation and angiogenesis of HBMECs in vitro through inhibition of PTEN. The angiogenic effects ofGSCs-derived exosomes overexpressing miR-26a in vivo were consistent with the above-mentioned in vitro findings.

Conclusion: Collectively, our study demonstrates that GSCs-derived exosomal miR-26a promotes angiogenesis ofHBMECs, highlighting an angiogenic role of miR-26a via exosomes.

Keywords: Glioma stem cells, Exosomes, microRNA-26a, Microvessel endothelial cells, Angiogenesis, Phosphatase andtensin homolog deleted on chromosome ten, PI3K/Akt signaling pathway

BackgroundGlioma is the general term used to indicate all types ofprimary central nervous system tumors and is the mostcommon type of primary brain tumors with poor prog-nosis [1, 2]. The incidence of glioma has been identi-fied to be 4.67–5.73 per 100,000 people [3]. Despitethe recent progress made in the therapeutic optionsfor glioma, the average survival time of patients withthe malignant glioma is less than 1.5 years [4]. Recur-rence of the gliomas is attributed to tumor-initiating

glioma stem cells (GSCs), which are essential for glio-magenesis and have been found to be intrinsicallyresistant to therapy [5]. As multipotent tumor-stimu-lating cells, GSCs exhibit stem cell features andexpress CD133 marker [6]. Moreover, GSCs have beendemonstrated to be involved in tumor growth andtumor angiogenesis [7, 8]. Along with a number ofother reasons, resistance to glioma therapies could po-tentially occur due to immortalized cellular growthand aberrant angiogenesis [9]. Angiogenesis is a hall-mark of cancer, due to its ability to facilitate cellstransforming from benign tumors to malignant ones[10]. Moreover, angiogenesis plays a significant role inthe progression and development of gliomas [11].

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

* Correspondence: wangzhifeidr@163.comDepartment of Neurosurgery, The Third Xiangya Hospital of Central SouthUniversity, No. 138, Tongzipo Road, Changsha 410013, Hunan Province,People’s Republic of China

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 https://doi.org/10.1186/s13046-019-1181-4

Therefore, extensive research is required to furtherelucidate the effects of GSCs on angiogenesis and theunderlying mechanism in human glioma.Extracellular vesicles (EVs) that mainly contains exo-

somes, ectosomes, microvesicles, microparticles, apop-totic bodies along with other EV subsets, could functionin the intercellular communication [12]. Exosomes are awell-studied subpopulation of small EVs [13, 14]. Exo-somes could be derived from different types of cellsincluding neurons, immune cells, stem cells and cancercells, which are considered to be one of the primary tar-gets of anti-cancer therapy due to their small sizes (40–100 nm) [15]. It has been well established that exosomesare capable of transferring proteins, messenger RNAs(mRNAs) and microRNAs (miRNAs) between cells [16].miRNA transfer mediated by exosomes has been recog-nized as a potential approach for miRNAs to exert theireffects on various biological functions in tumorigenesisand biogenesis [17]. Recent evidence has demonstratedthat exosomal miRNAs have been identified as promis-ing biomarkers for diagnosis and prognosis of glioblast-oma [18, 19]. miRNAs refer to small non-coding RNAmolecules, which serve as regulators of inhibition ofmRNA translation or reduction of mRNA stability bybinding to the 3′ untranslated region (3’UTR) of targetmRNA [20, 21]. In addition, the effects of miRNAs onglioma have been demonstrated with increasing clinicalimplications [22]. For instance, miR-26a has been re-ported to be amplified in glioma tissues and play a vitalrole in angiogenesis of glioma [23]. Similarly, miR-4467,miR-638 and miR-6727-5p are up-regulated in the exo-somes so as to affect the development of gliomas [24].Previously, Skog et al. have demonstrated that exosomesderived from GSCs can be internalized by brain micro-vascular endothelial cells (MVECs), thus serving as anew option for the delivery of genetic information andproteins to recipient cells in tumor environment [25]. Anumber of studies have shown that phosphatase andtensin homolog deleted on chromosome ten (PTEN) isone of the most common tumor suppressor genes inacti-vated in various cancers [26]. PTEN regulates variouscellular processes including cellular architecture, energymetabolism, proliferation and survival by mediating thePI3K-AKT-mTOR signaling pathway [26, 27]. Based onthe aforementioned findings, we hypothesized that miR-26a delivered from GSCs into MVECs via exosomescould potentially provide an effective therapeutic strat-egy for glioma by controlling PTEN and regulating thePI3K-Akt pathway. Therefore, in this study, we isolatedGSCs from glioma cells to elucidate the regulatory roleof exosomal miR-26a in angiogenesis of MVECs and themolecular mechanism associated with PTEN andPI3K-Akt pathway, in an attempt to provide a theoreticalfoundation for glioma treatment.

Materials and methodsEthics statementThis study was approved by the Ethics Committee ofThe Third Xiangya Hospital of Central South Universityand all the included patients in the study signed writtenconsents. All animal experiments were conducted instrict accordance with the Guidelines for Care and Useof Laboratory Animals.

Microarray analysisGlioma microarray datasets were obtained from theGene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), and the Limma packageof R language [28] was used to identify the differentiallyexpressed genes (DEGs) related to glioma. The Geneoncology (GO) enrichment analysis was conductedwith the WebGestalt database (http://www.webgestalt.org/option.php). Besides, the DisGeNET database(http://www.disgenet.org/web/DisGeNET/menu/home)was used to predict glioma-related genes. The inter-action among the DEGs was analyzed by using theSTRING database (https://string-db.org/), and thesignaling pathways related to PTEN was retrieved byusing the Kyoto Encyclopedia of Genes and Genomes(KEGG) database (https://www.kegg.jp/kegg/pathway.html). The miRDB database (http://mirdb.org/miRDB/index.html), mirDIP database (http://ophid.utoronto.ca/mirDIP/index.jsp), microT-CDS (http://diana.imis.athena-innovation.gr/DianaTools/index.php?r=microT_CDS/index) in DIANA database, and TargetScan data-base (http://www.targetscan.org/vert_71/) were used topredict the miRNAs that regulate PTEN.

Tissue collectionAll fresh tissue samples were surgically resected from46 patients with pathologically confirmed glioma be-tween September 2015 and September 2017 at the De-partment of Neurosurgery, The Third Xiangya Hospitalof Central South University. There were 26 males and20 females with an mean age of 42.16 ± 13.57 years(ranging from 18 to 72 years). Among these patients,there were 13 cases of astrocytoma I stage (WHO Istage), 16 cases of astrocytoma II stage (WHO II stage),13 cases of malignant anaplastic astrocytoma (WHO IIIstage), and 4 cases of glioblastoma multiforme (WHOIV stage) [29]. None of the patients included in thestudy had received radiotherapy or chemotherapy priorto the surgery. Subsequently, brain tissues were col-lected from 28 patients with acute craniocerebral injury(19 males and 9 females, 28–57 years, mean age of42.43 ± 8.14 years) that had received treatment of intra-cranial decompression.

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 2 of 15

Cell cultureThe glioma cell lines, SHG-44 (astrocytoma), BT325 (glio-blastoma), T98G (glioblastoma), A172 (glioblastoma), U251(astrocytoma), normal human glial cells (HEB), and humanbrain microvascular endothelial cells (HBMECs) were pur-chased from Cell Resource Center, Shanghai Institute forBiological Sciences, Chinese Academy of Sciences (Shang-hai, China). After recovery, the cells were cultured in a 5%CO2 and 37 °C incubator (Thermo Fisher Scientific, CA,USA) with saturated humidity with the Roswell ParkMemorial Institute (RPMI) 1640 culture medium (GibcoCompany, CA, USA) supplemented with 10% fetal bovineserum (FBS) (Gibco Company, CA, USA).

Cell transfectionThe HBMECs were seeded into a 12-well culture plate ata concentration of 1 × 105 cells/well 1 day before transfec-tion. When cell confluence reached 50–70%, HBMECswere transfected with miR-26a agomir, agomir-negativecontrol (NC), miR-26a antagomir, antagomir-NC, PTENoverexpression vector and PTEN overexpression NC(PTEN-NC) using lipo2000 (11,668,027, Thermo FisherScientific, CA, USA). After 6 h of culture, the culturemedium was renewed. And cells were collected for RNAand protein extraction after 48 h of transfection.

Dual-luciferase reporter gene assayThe biological prediction website Targetscan.org wasused to predict the binding sites of miR-26a in thePTEN sequence, and dual-luciferase reporter gene essaywas performed to verify whether PTEN was a direct tar-get gene of miR-26a. The synthesized PTEN 3’UTR genefragment was digested by enzyme sites XhoI and BamHIand inserted into pGL3-control vector (Promega Cor-poration, Madison, WI, USA). The mutant (Mut) siteswere designed in the supplementary sequences of seedsequences in PTEN wild type (Wt) and PTEN Mut se-quences, Xho I and BamH I. The target fragments wereinserted into the pGL3-control vectors using T4 DNAligase. The pGL3-PTEN-Wt and pGL3-PTEN-Mut wereco-transfected with miR-26a mimic and NC sequencesof miR-26a mimic into glioma cell U251 (Cell ResourceCenter, Shanghai Institute for Biological Sciences,Chinese Academy of Sciences, Shanghai, China), respect-ively. After a 48-h transfection, the cells were collectedand lysed. Dual-Luciferase® Reporter Assay System kit(Promega Corporation, Madison, WI, USA), and Lumin-ometer TD-20/20 (E5311, Promega Corporation, Madi-son, WI, USA) were employed to determine theluciferase activity.

RNA isolation and quantitationThe total RNA was extracted by using a Trizol kit (Invitro-gen Inc., Carlsbad, CA, USA) and reversely transcribed into

complementary DNA (cDNA). The primer sequences ofmiR-26a and PTEN were designed and synthetized byAOKE Biotechnology Co., Ltd. (Zhenjiang, China) (Table 1).RT-qPCR was conducted with the use of the ABI StepOne-Plus real time quantitative PCR instrument (StepOnePlus,ABI Company, Oyster Bay, NY, USA). U6 was used as theinternal reference of miR-26a, while glyceraldehyde-3-phos-phate dehydrogenase (GAPDH) was used as the internal ref-erence of other genes. The relative mRNA expression of thetarget gene was analyzed by 2-ΔΔCt method, using the follow-ing formula: △△Ct =△Ct experimental group - △Ct control group,△Ct =Ct target gene - Ct internal reference.

Western blot analysisThe total protein was extracted by using radio-immuno-precipitation assay (RIPA) lysis buffer (BB-3209, BestBioScience, Shanghai, China). The proteins were separatedusing sodium dodecyl sulfate (SDS)-polyacrylamide gelelectrophoresis (PAGE), and transferred onto a polyviny-lidene fluoride (PVDF) membrane with constant voltageof 80 V. After the membranes were blocked for 1 h, theywere incubated with diluted primary antibodies, rabbitanti-human PTEN (1: 100, ab32199; Abcam Inc.. Cam-bridge, MA, USA), vascular endothelial growth factor(VEGF) (1: 100, sc4570, Santa Cruz Biotechnology, Inc.,Santa Cruz, CA, USA), matrix metalloproteinase(MMP)-2 (1: 100, ab37150, Abcam Inc.. Cambridge,MA, USA), MMP-9 (1: 100, ab73734, Abcam Inc.. Cam-bridge, MA, USA), protein kinase B (Akt) (1: 100,ab8805, Abcam Inc.. Cambridge, MA, USA), p-Akt (1:100, ab192623, Abcam Inc.. Cambridge, MA, USA), andGAPDH (1: 100, ab9385, Abcam Inc.. Cambridge, MA,USA) (internal reference) for 1 h at 37 °C. TBST washes(5 min × 3) were followed. Horseradish peroxidase(HRP)-labeled rabbit anti-human IgG (1: 20000,ab205718, Abcam Inc.. Cambridge, MA, USA) served asthe secondary antibody. The membrane was developedby enhanced chemiluminescence (ECL).

Table 1 Primer sequences of related genes for RT-qPCR

Gene Sequence

miR-26a F: 5′-GCCTAACCCAAGAAGGGAAA-3′

R: 5′-TCCCTCTCATCTGGACAACC-3′

PTEN F: 5′-AAGAGGCCCTAGATTTCTATG-3′

R: 5′-CAGTAGAGGAGCCGTCAAAT-3′

U6 F: 5′-AATTGGAACGATACAGAGAAGATTAGC-3′

R: 5′-TATGGAACGCTTCACGAATTTG-3′

GAPDH F: 5′-GAGAGACCCCACTTGCTGCCA-3′

R: 5′-GGAAGAAGTTCCCATCGTCA-3′

RT-qPCR reverse transcription quantitative polymerase chain reaction, miR-26amicroRNA-26a, PTEN phosphatase and tensin homolog deleted onchromosome ten, GAPDH glyceraldehyde-3-phosphate dehydrogenase

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 3 of 15

GSC isolation and characterizationThe U251 cells in logarithmic growth phase were col-lected in an aseptic centrifuge tube containing Neurobasalstem cell culture medium. The immunomagnetic beadsorting kit and instrument were used for the separation ofCD133+ cells in accordance with the instructions [30].Cells were added with immunomagnetic bead sorting so-lution (200 μL/108 cells), and triturated into single cellsuspension. Meanwhile, the CD133 antibody-beads com-plex was incubated together with cells (100 μL/105 cells)at 4 °C for 30min, and the sorting solution was added torinse cells (1mL/108 cells). Following centrifugation andthe removal of the cell supernatant, the cells were col-lected and re-suspended with the addition of sorting solu-tion (500 μL/108 cells). Afterwards, the cell suspensionwas added into the separation column, followed by theaddition of 2mL sorting solution for elution of CD133+

cells. The separated CD133+ and CD133− cells were addedwith 50 μL CD133/2 (293C3)-PE antibody or 50 μL homo-typic control IgG2b-PE antibody, respectively, and sortedby flow cytometry.

GSC transfectionA day prior to transfection, the cells were incubated withDulbecco’s modified Eagle medium (DMEM)/F12 culturemedium without penicillin-streptomycin at a density of1 × 105 cells/well. When GSC confluence reached 70–90%, the cells were transfected in strict accordance withthe instructions of FuGENE HD6 transfection kit (E2311,Promega Corporation, Beijing, China). The GSCs weretransfected with miRNA agomir (GSCmiR-26a agomir) andagomir NC (GSCagomir-NC), miRNA antagomir(GSCmiR-26a antagomir) and antagomir NC (GSCantagomir-NC),respectively. The final concentration of miRNA agomir orantagomir was 100 nmol/L. After a 6-h transfection, themedium was replaced by fresh stem cell medium. After48 h of transfection, the cells obtained were used for sub-sequent experiments.

GSC exosome extractionThe transfected GSCs were seeded in the exosomes-de-pleted RPMI 1640 culture medium containing 10% FBS,and cultured in an incubator with 5% CO2 at 37 °C.After 3 days, the cell supernatant was collected and cen-trifugation was carried out to remove cell debris. Theexosomes were extracted on the basis of the instructionsof Hieff™ Quick exosome isolation kit (41201ES50, Yea-sen Company, Shanghai, China). The cell supernatantand separation reagent of exosome were added into thecentrifuge tube (EP tube) at a ratio of 2: 1 and incubatedat 4 °C overnight. Next, the samples were centrifuged at10000×g at 4 °C for 1–2 h, followed by the removal ofthe supernatant, and the collection of the precipitates(exosomes). In accordance with volume ratio of initial

culture medium and resuspension (10: 1), the sampleswere re-suspended in phosphate buffered saline (PBS),after which 30 μL of the re-suspended exosomes wereplaced in an EP tube and added with the equal volumeof RIPA lysis buffer, mixed, and placed on ice. The exo-somes were lysed twice using microwave method for 10s. Subsequently, protein concentration in exosomes wasmeasured via bicinchoninic acid (BCA) quantitative kit(Beyotime Biotechnology, Jiangsu, China). The particlesize distribution of exosomes was analyzed with the useof a Nanosight nanoparticle tracking analyzer (MalvernPanalytical, Malvern, UK). Flow cytometry was per-formed to determine exosome surface markers (CD63and CD81). Morphology of exosomes was characterizedunder a transmission electron microscope (TEM)(JEM-1010, JEOL, Tokyo, Japan). Finally, the exosomessecreted from GSCs (GSCs-exosomes) were identified.The exosomes were immediately fixed in 2.5% glutar-

aldehyde at 4 °C. After fixation, the exosomes were dehy-drated by gradient alcohol and embedded with epoxyresin. Ultrathin sections were stained with uranyl acetateand lead citrate and observed under a TEM (JEM-1010,JEOL, Tokyo, Japan).

Flow cytometryOne μL beads were added into the EP tube containing300 μL morpholine-ethanesulfonate (MES) buffer, andevenly mixed. Next, the mixture were centrifuged at1610×g for 5 min. The supernatant was discarded byusing a pipette. The beads were washed and then sus-pended using 300 μL MES buffer. Next, GSCs-exosomesuspension were incubated together with the beads atroom temperature for 1 h. The EP tube was rotated at 4°C overnight. On the following day, the EP tube was cen-trifuged at a low speed for 1 min, and the samples werediluted into a final concentration of 200 μmol/mL withthe addition of 1 mol/L glycine. Following an incubationperiod of 30 min at room temperature, the samples werere-suspended with 300 μL of 3% FBS-MES buffer and in-cubated with antibodies anti-CD63 (FITC) (ab18235,Abcam Inc., Cambridge, MA, USA) and anti-CD81(FITC) (ab239256, Abcam Inc., Cambridge, MA, USA)at room temperature for 40 min. Subsequently, the sam-ples were washed twice with 300 μL of 3% FBS-MESbuffer. The samples were centrifuged at 16102×g for 5min to remove the cell supernatant. Finally, the sampleswere re-suspended with 250 μL PBS and subjected toflow cytometric detection.

Co-culture of exosomes and HBMECsThe HBMECs in logarithmic growth phase were seededinto a 6-well plate of RPMI 1640 culture medium con-taining 5% FBS (5 × 105 cells/well). Subsequently, 20 μgexosomes extracted from transfected GSCs were mixed

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 4 of 15

evenly with PKH26 (RED, 1: 1000), and incubated at 37 °C for 15 min. Finally, PKH26-traced exosomes wereco-cultured with green fluorescent protein (GFP)-labeledHBMECs, followed by incubation with 200 μL PBS con-taining 1% BSA at room temperature for 20 min. Thesamples were stained with 4′,6-diamidino-2-phenylin-dole (DAPI) to stain nuclei in blue and VECTASHIELDmounting media (Vector Labs, CA, USA). The OlympusBX61 confocal fluorescence microscope (Zeiss Meta 510,Thornwood, NY, USA) was adopted to observe internal-ization of exosomes by HBMECs.

Enzyme-linked immunosorbent assay (ELISA)A total of 50mg of tissues were homogenized for 2min,and centrifuged at 1610×g and 4 °C for 10min, after whichthe cell supernatant was obtained. VEGF concentration inthe supernatant was measured in accordance with theinstructions of the ELISA kit (69–50,049, Wuhan MSK Bio-technology Co., Ltd., Wuhan, China). The absorbance wasdetected in microplate reader (Synergy 2, BioTek, USA) atthe wavelength of 450 nm within 3min.

5-ethynyl-2′-deoxyuridine (EdU) labelingHBMECs in logarithmic growth phase that had beenco-cultured with GSCs-exosomes were seeded into a96-well plate (3 × 104 cells/well). The cells were cul-tured with EdU culture medium (100 μL/well) for 2 h,followed by incubation with 100 μL fixative for 30 min.Subsequently, incubation was carried out in each wellwith 2 mg/mL glycine for 5 min and with 100 μL of0.5% TritonX-100 in PBS for 10 min. The cells were in-cubated again with 1 × Apollo dyeing reaction solutionwithout light exposure for 30 min, and incubated with100 μL 1 × Hoechst 33342 reaction solution withoutlight exposure for 30 min in successive. After staining,100 μL anti-fluorescence quenching agent was addedinto each well. A total of 6–10 fields were selected foreach well, which were observed and photographedunder the fluorescence microscope.

Transwell assayAfter 48 h of transfection, the Matrigel (YB356234,Shanghai YU BO Biological Technology Co., Ltd.,Shanghai, China) preserved at − 80 °C was dissolved at4 °C overnight. The cells were then trypsinized, andprepared into cell suspension. Each well in the apicalchamber was added with cell suspension (1 × 104 cells/μL), and the basolateral chamber was added with800 μL medium containing 20% FBS. Following an in-cubation period of 16 h at 37 °C, the Transwell chamberwas fixed with formaldehyde for 10 min, and cleanedthree times with water. Subsequently, the samples werestained with 0.1% crystal violet for 30 min. The cells on

the surface were wiped off by cotton ball. After 24 h,the cells were observed, photographed, and countedunder an inverted microscope.

Tube formation assayCalcein-labeled HBMECs were seeded into a 24-wellplate at the density of 2 × 105 cells/mL. After cells at-tached to the wall of well, exosomes from GSCs whichwere transfected with miR-26a agomir or miR-26aantagomir were incubated at 37 °C with 5% CO2 for 12–24 h. Tube formation was observed under the confocalmicroscope. The images were captured at a fixed time.The number and length of branches were measured, re-corded, and statistically analyzed [31].

Angiogenesis assay on chicken chorioallantoic membrane(CAM)A total of 60 embryonated chicken embryos were incu-bated at (37.5 ± 0.5)°C with 65% humidity. On the 3rdday after incubation, the vascular network growth wasexamined by candling, with 48 well-developed ones en-rolled. GSCs transfected with agomir-NC, miR-26aagomir, antagomir-NC, or miR-26a antagomir wereinjected into the chicken embryos (n = 12 per group).On the 7th day of incubation, the fertilized chicken em-bryos were disinfected with ethanol. A small window(1.5 cm × 1.5 cm) was excised at the lower right side ofthe embryo head to expose the CAM. The sample filterpaper was placed on CAM and yolk sac membrane withless blood vessels. After incubation for 10 days in athermostat box, the transparent adhesive tape coveredon the chicken embryos was removed. The angiogenesiswas observed under the microscope and photographedby a digital camera [32].

Tumorigenicity assay in nude miceA total of 24 BALB/c female nude mice (aged 4–6 weeksold and weighing 16–22 g) purchased from Shanghai La-boratory Animal Center, Chinese Academy of Science wererandomly grouped into the GSCagomir-NC, GSCmiR-26a agomir,GSCantagomir-NC, and GSCmiR-26a antagomir groups, with 6mice in each group. Subsequently, the nude mice wereinjected with the transfected GSCs. The single cell suspen-sion was subcutaneously injected into the back of the nudemice using a disposable sterile syringe (4 × 106 cells/mL).The tumor volume was calculated as V =W2 × L × 0.52[33]. After 35 days, nude mice were euthanized, followedby the resection of the implanted tumors, which were fixedin 4% neutral formaldehyde overnight. After being embed-ded with paraffin, the samples were cut into 5 μm sections.

Microvessel density (MVD) detectionThe streptavidin-peroxidase immunohistochemicalstaining was performed with CD31 (1: 200) primary

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 5 of 15

antibody. The primary antibody replaced with PBS inNC, and the CD31-positive sections served as positivecontrol. Five fields under the microscope for each sec-tion were randomly selected, and 100 cells in eachfield were counted. The MVD value was expressed bythe number of newly formed vessels (CD31 positivestaining cells).

Statistical analysisStatistical analysis was performed by using the SPSS21.0 software (IBM Corp. Armonk, NY, USA), with nor-mality and homogeneity of variance tested. The datawith normal distribution were expressed as mean ±standard deviation (SD). Comparisons between gliomabrain tissues and control brain tissues were analyzed byusing unpaired t-test, while comparisons among mul-tiple groups were analyzed by using one-way analysis ofvariance (ANOVA) with Tukey’s post hoc test. The dataat different time points were analyzed by repeatedmeasurement ANOVA, and the correlation analysis wasconducted using Pearson’s correlation analysis. p < 0.05was considered as statistically significant.

ResultsPTEN and miR-26a are involved in gliomaThe expression profile of glioma (GSE50161) was ob-tained from the GEO database and analyzed in order toinvestigate DEGs in glioma. The results found 4935DEGs in total. GO enrichment analysis further showedthat these DEGs were mainly enriched in three categor-ies, including “biological regulation”, “membrane” and“protein binding” (Fig. 1a). In order to further clarifythe DEGs related to glioma, interaction analysis andgene interaction network were performed on the top 50DEGs. The known genes related to glioma were ob-tained from DisGeNET database, and the genes withscores > 0.2 were intersected with DEGs, which identi-fied 26 intersected genes. Further gene interaction ana-lysis was conducted on these 26 genes (Fig. 1b). Amongthem, PTEN, a tumor-suppressor gene, was located atthe core position (Fig. 1c). Moreover, the PTEN-relatedsignaling pathways were further retrieved from theKEGG database. Based on the findings, PTEN couldpotentially regulate the PI3K/Akt signaling pathway(map 05214). PTEN was down-regulated in the gliomaexpression dataset (Fig. 1d), indicating that PTEN mayact as regulator in glioma via the PI3K/Akt signalingpathway. According to recent studies, miRNAs such asmiR-29 and miR-19a-3p regulate PTEN to function astumor promoter or suppressor [34, 35]. Four predictionwebsites (miRDB, mirDIP, DIANA and TargetScandatabase) were employed to predict the miRNAs thatcould regulate PTEN in order to further investigate theupstream regulation mechanism of PTEN. Meanwhile,

glioma-related miRNA expression dataset GSE90604was obtained from the GEO database and the up-regu-lated miRNAs were intersected with the putative miR-NAs that could regulate PTEN. As shown in Fig. 1e,there were five miRNAs (hsa-miR-92b-3p, hsa-miR-19a-3p, hsa-miR-19b-3p, hsa-miR-4429 and hsa-miR-26a) in the intersection. Moreover, further analysis onthese miRNAs found that compared with control braintissues, there was a significant increase in miR-26a inglioma tissues (Fig. 1f ). The above findings suggestedthat PTEN is lowly expressed while miR-26a highlyexpressed in glioma, and miR-26a might target PTEN.

miR-26a is up-regulated in both glioma tissues and GSCsTo determine whether miR-26a was expressed in glioma tis-sues, its expression was evaluated in control brain tissuesand glioma tissues by RT-qPCR, respectively (Fig. 2a). Com-pared with normal human glial cells (HEB), miR-26a wassignificantly up-regulated in five glioma cell lines, includingSHG-44 (astrocytoma), BT325 (glioblastoma), T98G (glio-blastoma), A172 (glioblastoma) and U251 (astrocytoma).Among those cell lines, U251 cell line presented with thehighest expression of miR-26a. GSCs are a kind of multi-functional tumor initiating cells with stem cell properties[36, 37] and express CD133 marker [38, 39]. To examinewhether miR-26a was also highly expressed in GSCs, flowcytometry was adopted to separate GSCs from U251 cells,and the CD133+ cellular expression was determined usingflow cytometry/immunomagnetic beads (Fig. 2b). The suc-cessful separation of the GSCs was indicated by the evidentincrease observed in the CD133+ cellular expression. Inaddition, as shown in Fig. 2c, miR-26a expression was obvi-ously elevated in GSCs. These findings suggested thatmiR-26a was up-regulated in both glioma tissues and GSCs.

GSCs transfer miR-26a to HBMECs through exosomesThe molecular mechanisms of GSCs in regulatingHBMECs were investigated, the exosomes successfullyisolated from GSCs (Additional file 1: Figure S1) wereco-cultured with HBMECs. RT-qPCR was conducted tomeasure the expression of miR-26a and PTEN. Based onthe results, miR-26a was up-regulated both in GSCstransfected with miR-26a agomir and exosomes derivedfrom GSCmiR-26a agomir compared to that in agomir-NC-transfected GSCs and GSCagomir-NC derived exosomes,respectively (Fig. 3a-b). Notably, the overexpression ofmiR-26a, and down-regulation of PTEN were observedin HBMECs co-cultured with GSC-exosomes in theevent that miR-26a agomir was transfected into GSCsbefore co-culture (Fig. 3b). Notably, GSC-exosomes wereobserved to be internalized by HBMECs and distributedaround the nucleus under the confocal microscope(Fig. 3c). Therefore, these findings led to the conclusion

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 6 of 15

that GSCs transferred miR-26a to HBMECs throughexosomes.

Up-regulation of miR-26a promotes proliferation,migration and angiogenesis of HBMECsTo determine whether miR-26a could affect prolifera-tion, migration and angiogenesis of HBMECs in vitro,the HBMECs were transfected with miR-26a agomir, or

miR-26a antagomir. As shown in Fig. 4a-c, the overex-pression of miR-26a enhanced the HBMEC proliferation,migration and tube formation. In contrast, inhibition ofmiR-26a expression caused the reduction in HBMECproliferation, migration and tube formation. VEGF ex-erts strong effects on promoting angiogenesis whencombined with corresponding receptors [40]. MMP is atype of extracellular proteolytic enzyme, which can

PLEKHH2CA1

EDNRAUSP3

LOX NKIRAS1

PPM1K

NAP1L2

C1orf112

RNF213

GAS6PTEN

HERC5SIRPA

SRPX2

IQGAP3

C2orf68

RBCK1

SAMD9LAEBP1TRIM21

SPTBN2

PSD3

LY96

PAPD5

P4HTM

PTPRNWDR12

DSCC1

LAPTM4B

IGSF8 PTGFRN

UNC79SYT9

STON1

Normal Tumor

62

34

14

1

2

221

21

00

0

50

70

20

86

68

8

175

17

1

13

0

0

0

0

0

7 3 27

miR

DB

DIANA

mirD

IP

TargetScan

GSE90604

A

B

E

C

9.5

10.0

10.5

11.0

11.5

12.0

Exp

ress

ion

PT

EN

(GS

E50

161)

2.07E−06D

Rel

ativ

e ex

pres

sion

leve

l

miR-92b-3p miR-19a-3p miR-19b-3p miR-26a0

2

4

6

8Control brain tissueGlioma tissue

F

4909 5226

GSE50161

DisGeNET

miR-4429

****

*

Fig. 1 PTEN is poorly expressed in glioma and predicted to be a target of miR-26a. a GO enrichment analysis of DEGs in expression dataset of glioma(GSE50161); three histograms refer to the biological process (BP), cellular component (CC) and molecular function (MF) enrichment, respectively; theabscissa refers to GO entry, while the ordinate refers to the number of DEGs; b the intersection of DEGs and the known genes related to glioma. Twocircles represent DEGs obtained from GSE50161 datasets (blue) and the known genes related to glioma obtained from DisGeNET (red), and the middlepart indicates their intersection; c interaction analysis of glioma DEGs; each circle represents one gene, the line among circles represents the interactionamong genes, and the color of circle represents the core degree of gene in the whole network; d PTEN expression in GSE50161 dataset; the abscissa refersto sample type, and the ordinate refers to gene expression level; the left box plot refers to the normal group, and the right box plot refers to the gliomagroup; e putative miRNAs that regulate PTEN; the intersection of putative miRNAs targeting PTEN obtained from the four databases and up-regulatedmiRNAs in glioma from GSE90604 dataset; the middle part refers to the intersections among five datasets; f the expression of miR-92b-3p, miR-19a-3p,miR-19b-3p, hsa-miR-4429 and miR-26a in control brain tissues and glioma tissues determined by RT-qPCR; the data between control brain tissues (n= 28)and glioma tissues (n= 46) are compared by unpaired-t test. The experiments are repeated 3 times. * p< 0.05, compared with the control group. DEGs,differentially expressed genes; miR/miRNA, microRNA; GO, gene oncology; PTEN, phosphatase and tensin homolog deleted on chromosome ten

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 7 of 15

degrade basal membrane and promote invasion and mi-gration of malignant tumors, of which MMP-2 andMMP-9 are two primary enzymes to degrade type IVcollagen [41]. The results of Western blot analysis(Fig. 4d) revealed that the up-regulation of miR-26a ele-vated the protein levels of VEGF, MMP-2, and MMP-9,whereas inhibition of miR-26a expression decreased pro-tein levels of VEGF, MMP-2, and MMP-9. These find-ings suggested that elevated miR-26a promoted HBMECproliferation, migration and tube formation in vitro.

miR-26a activates the PI3K/Akt signaling pathway bydown-regulating PTEN in gliomaTo further clarify the the regulatory effect of miR-26aon PTEN, the miR-26a binding sites in PTEN werepredicted using the biological prediction website. Asshown in Fig. 5a, PTEN 3’UTR contained potentialmiR-26a binding sites. In addition, dual-luciferase re-porter gene assay was performed with PTEN-Wt andPTEN-Mut co-transfected with miR-26a mimic or NCinto glioma cells. Comparison with the NC mimic, theluciferase activity of PTEN-Wt was inhibited in thepresence of miR-26a mimic (p < 0.05) (Fig. 5b), suggest-ing that miR-26a could specifically bind to PTEN. Fur-thermore, the PTEN expression in the control braintissue and glioma tissue was quantified by RT-qPCR(Fig. 5c-d). The expression of PTEN was reduced in gli-oma tissues compared to the control brain tissue, and itwas negatively related to miR-26a expression. Consist-ently, the results from the Western blot analysis in

HBMECs revealed that PTEN protein level was reducedin the presence of miR-26a agomir (Fig. 5e). In con-trast, PTEN protein level was increased in the presenceof miR-26a antagomir, suggesting that miR-26a couldsuppress the PTEN expression. To further investigatewhether miR-26a targets PTEN to regulate the PI3K/Akt signaling pathway in glioma, phosphorylation ofproteins related to the PI3K/Akt signaling pathway wasmeasured by Western blot analysis. As shown in Fig. 5f,p-Akt/Akt ratio was reduced in presence of PTEN, butincreased in miR-26a agomir. Notably, p-Akt/Akt ratiowas significantly decreased in HBMECs following theup-regulation of both miR-26a and PTEN as comparedto that after the up-regulation of miR-26a alone. Basedon the aforementioned results, miR-26a activated thePI3K/Akt signaling pathway through the down-regula-tion of PTEN.

Exosomal miR-26a promotes proliferation andangiogenesis of HBMECsThe exosomes isolated from transfected GSCs wereco-cultured with HBMECs in order to determinewhether exosomal miR-26a mediates PTEN-mediatedPI3K/Akt signaling pathway to regulate proliferation andangiogenesis of HBMECs. According to the ELISA ana-lysis, VEGF level was increased in HBMECs co-culturedwith exosomes overexpressing miR-26a (Fig. 6a). More-over, RT-qPCR showed that PTEN expression was de-creased in HBMECs co-cultured with exosomes

A

Rel

ativ

e ex

pres

sion

of m

iR-2

6a

HEB

SHG-44

BT325

T98G

A172

U251

0

1

2

3

4

5

**

**

*

B

Rel

ativ

e ex

pres

sion

of m

iR-2

6a

NGSCs GSCs0

2

4

6 *

C

CD

133+

cel

ls (

%)

U251 cells GSCs0

20

40

60

80

100*

U251 cells

14.3%

GSCs

89.6%

101

004002

0

tnuoC

FITC-conjugated CD133

006

102 103 104 105 106 107 101

0004

002

tnuoC

FITC-conjugated CD133

006

102 103 104 105 106 107

Fig. 2 miR-26a is highly expressed in both glioma tissues and GSCs. a miR-26a expression in normal human glial cells (HEB) and glioma cell lines (SHG-44, BT325, T98G, A172, and U251); b isolation of GSCs from U251 cells using flow cytometry; c miR-26a expression in GSCs and NGSCs; *, p< 0.05,compared with the HEB (a), U251 cells (b), or NGSCs (c); all data were expressed as mean ± standard deviation; comparisons among multiple groupswere analyzed by the one-way ANOVA; the experiment was repeated three times; ANOVA, analysis of variance; miR-26a, microRNA-26a; GSCs, gliomastem cells; NGSCs, non-glioma stem cells

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 8 of 15

overexpressing miR-26a (Fig. 6b). Consistently, theWestern blot analysis revealed that restoration of PTENcaused an evident decrease in the p-Akt/Akt ratio inHBMECs co-cultured with exosomes overexpressingmiR-26a (Fig. 6c). Moreover, as shown in Fig. 6d-f,HBMECs co-cultured with exosomes with overexpressedmiR-26a presented with enhanced HBMEC proliferation,migration, and tube formation ability. The results fromthe Western blot analysis also showed that protein levelsof VEGF, MMP-2, and MMP-9 were significantly in-creased in HBMECs co-cultured with exosomes overex-pressing miR-26a (Fig. 6g). Therefore, we concluded thatexosomal miR-26a inhibited the PTEN expression andactivated the PI3K/Akt signaling pathway to promoteproliferation and angiogenesis of HBMECs.

miR-26a overexpression in GSCs promotes angiogenesisin vivoThe CAM models were established to investigate the ef-fect of miR-26a on angiogenesis of HBMECs in vivo. Asshown in Fig. 7a, the branches of blood vessels (5mm)around the CAM vehicle were evidently increased, andthe angiogenesis was enhanced in response to

overexpressed miR-26a. In contrast, when miR-26a ex-pression was inhibited, the branches of blood vessels (5mm) around the CAM vehicle obviously reduced, andangiogenesis was inhibited. In order to further confirmthat GSCs influenced angiogenesis of HBMECs in vivo,the GSCs transfected with miR-26a agomir, miR-26aantagomir and their corresponding NC plasmids wereinjected into the nude mice to observe the tumorigenicityand metastasis in vivo. The tumorigenicity of GSCs weretested in nude mice. The tumor size and weight were in-creased by enhancement of miR-26a expression, but de-creased by inhibition of miR-26a (Fig. 7b-c). ELISA data(Fig. 7d) showed that there was an evident increase inserum VEGF level in the nude mice when miR-26a ex-pression was enhanced, but significantly reduced whenmiR-26a expression was inhibited. Western blot analysis(Fig. 7e) showed that miR-26a overexpression led to thereduction of PTEN protein level and elevation in VEGFprotein level, while these were reversed following the in-hibition of miR-26a expression. Furthermore, the resultsfrom the immunohistochemistry (Fig. 7f) indicated thatthe angiogenesis endothelial marker, CD31-MVD valuewas markedly elevated in the event that miR-26a

A

C

B

Fig. 3 GSCs deliver miR-26a to HBMECs through exosomes. a miR-26a expression in GSCs transfected with miR-26a agomir and their derived exosomesusing RT-qPCR; bmiR-26a expression and PTEN expression in HBMECs co-cultured with exosomes derived from miR-26a agomir-transfected GSCs usingRT-qPCR; c exosomes and uptake of exosomes into HBMECs observed under the confocal microscope (× 400); exosomes labeled with PKH26 was stainedred; GFP-labeled HBMECs were stained green; the nucleus stained with DAPI was blue; *, p< 0.05, compared with the GSCagomir-NC or control group; alldata were expressed as mean ± standard deviation; comparisons between two groups were analyzed using unpaired t-test; comparisons among multiplegroups were analyzed by the one-way ANOVA; the experiment was repeated three times; ANOVA, analysis of variance; RT-qPCR, reverse transcriptionquantitative polymerase chain reaction; miR-26a, microRNA-26a; GFP, green fluorescent protein; PTEN, phosphatase and tensin homolog deleted onchromosome ten; HBMECs, microvascular endothelial cells; GSCs, glioma stem cells; DAPI, 4′,6-diamidino-2-phenylindole; NC, negative control

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 9 of 15

expression was increased. Taken together, the above find-ings demonstrated that GSCs overexpressing miR-26acould potentially enhance angiogenesis of HBMECs andtumor growth of nude mice, down-regulate PTENexpression and up-regulate VEGF expression.

DiscussionGlioma is a type of malignant brain tumor that arisesfrom the glial cells of the brain, ranking as one of themost commonly occurring intracranial tumors [4].GSCs are tumor cells found in primary glioblastomasthat possess stem-cell-like properties, such as self-

renewal ability and further capacity to generate het-erogeneous offspring [42]. It has been suggested thatnanoscale vesicles, exosomes derived from GSCsserved as mediators of cell-cell communication tocontrol tumor immunity in glioma [43]. Exosomes de-rived from mesenchymal stem cells leads to the accel-eration of angiogenesis by transferring miRNAs toendothelial cells [44]. In the present study, we showedthat GSCs-exosomes overexpressing miR-26a couldlead to the inhibition of PTEN and activation of thePI3K/Akt signaling pathway to promote the angiogen-esis of HBMECs.

Pos

itive

exp

ress

ion

of E

du r

ate

(%)

0

20

40

60

80

*

#

A

B

C

D

Tub

e le

ngth

(µm

)

12 h 24 h

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

Num

ber

of b

ranc

hes

12 h 24 h0

50

100

150

**

#

#

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

Rel

ativ

e pr

otei

n ex

pres

sion

VEGF MMP-2 MMP-90.0

0.5

1.0

1.5

2.0

2.5

*

*

*

##

#

Cel

l num

ber

of m

igra

tion

0

50

100

150

*

#

24 h

12 h

agomir-NC antagomir-NCmiR-26a agomir miR-26a antagomir

agomir-NC antagomir-NCmiR-26a agomir miR-26a antagomir

agomir-NC antagomir-NCmiR-26a agomir miR-26a antagomir

agom

ir-NC

miR

-26a

ago

mir

anta

gom

ir-NC

miR

-26a

ant

agom

ir

VEGF

MMP-2

MMP-9

GAPDH

25 µm 25 µm 25 µm 25 µm

25 µm 25 µm 25 µm 25 µm

25 µm 25 µm 25 µm 25 µm

25 µm 25 µm 25 µm 25 µm0

200

400

600

800

1000

*

*

#

#

Fig. 4 Restored miR-26a enhances HBMEC proliferation, migration, tube formation, and angiogenesis. a proliferation of HBMECs transfected with miR-26aagomir or miR-26a antagomir detected using EdU assay (× 400); b migration of HBMEC transfected with miR-26a agomir or miR-26a antagomir assessedusing transwell assay (× 400); c tube formation of HBMECs transfected with miR-26a agomir or miR-26a antagomir (× 400); d protein levels of VEGF, MMP-2and MMP-9 of HBMEC transfected with miR-26a agomir or miR-26a antagomir measured using Western blot analysis; *, p< 0.05, compared with theagomir-NC group; #, p< 0.05, compared with the antagomir-NC group; all data were expressed as mean ± standard deviation; comparisons between twogroups were analyzed using unpaired t-test; comparisons among multiple groups were analyzed by the one-way ANOVA; the experiment was repeatedthree times; ANOVA, analysis of variance; miR-26a, microRNA-26a; HBMECs, human brain microvascular endothelial cells; EdU, 5-ethynyl-2′-deoxyuridine;VEGF, vascular endothelial growth factor; MMP-2, matrix metalloproteinase-2; MMP-9, matrix metalloproteinase-9; NC, negative control

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 10 of 15

Exosomes are involved in several cellular processesincluding intracellular communication, material trans-portation on the surface of cell membrane and theregulation of cellular content during membrane fusion[45]. Exosomes derived from glioblastoma tumor cellscontain intact mRNA that can be transferred to recipi-ent cells and translated into functional proteins [46]. Astudy by Brower et al. demonstrated that miRNAs,which are dysregulated in malignant cells, are impli-cated in tumorigenesis [22]. Exosomes play a criticalrole in cellular communication by directly transferringgenetic materials including mRNA and miRNA be-tween cells [47, 48]. An example of this can be that theexosomes found in gliomas could deliver miR-1 to re-cipient cells, which then affects glioma cell invasionand proliferation as well as endothelial cell tube forma-tion [49]. Moreover, overexpressed miR-21 in GSCs-de-rived exosomes can be delivered to endothelial cells(ECs) and promote the angiogenic ability of ECs [6].Consistently, our results showed that miR-26a wasup-regulated in glioma tissues and GSCs. Moreover,GSCs-derived exosomes were observed to transfermiR-26a in order to favor for elevating its expression. Arecent study has reported that miR-26a contributes toenhanced angiogenesis in glioma by binding to prohibi-tin [23]. Moreover, miR-26a overexpression contributedto the enhanced proliferation, migration and tube for-mation of HBMECs. Consistent with our finding, highexpression of miR-26a in glioma tissues contributes tothe tumorigenesis of glioma [50]. Interestingly, miR-26a

transferred from GSCs-derived exosomes into HBMECshad the ability to induce the inhibition of PTENexpression.The results from the dual-luciferase reporter assay

revealed that miR-26a targeted PTEN, and down-regu-lated PTEN expression. And PTEN was found to bedown-regulated in GSCs as well. A previously con-ducted study demonstrated that miR-26a targets PTENto promote the metastatic capacity of lung cancer cells[51]. In addition, the dysregulation of miR-26a has beenreported to affect cell proliferation in glioblastoma bybinding to PTEN [52]. miR-26a has also been found tobe up-regulated in glioma tissues and it contributes tothe tumorigenesis of glioma by targeting PTEN [50].Consistently, our findings suggested that exosomes de-rived from GSCs with up-regulated miR-26a could pro-mote proliferation, migration and tube formation ofHBMECs in vitro. A previous study has indicated thatmiR-93 plays a regulatory role in glioma progressionthrough the suppression of PTEN by activating thePI3K/Akt signaling pathway [53, 54]. The up-regulationof miR-26a in GSCs results in an increase in the phos-phorylation of Akt, suggesting that the PI3K/AKT sig-naling pathway is activated by miR-26a in glioma. Theaforementioned findings from our study were furthersupported by the findings reported by Cai et al., whichdemonstrated that the activation of the PI3K/Akt sig-naling pathway leads to the enhancement of prolifera-tion, migration and tube formation of endothelial cellsco-cultured with glioma in vitro [55].

Rel

ativ

e ex

pres

sion

of P

TE

N p

rote

in

agomir-NCmiR-26a agomirantagomir-NCmiR-26a antagomir

agom

ir-NC

miR

-26a

ago

mir

anta

gom

ir-NC

miR

-26a

ant

agom

ir

0.0

0.5

1.0

1.5

2.0

2.5

*

#

E

PTEN

GAPDH

Rel

ativ

e ex

pres

sion

of p

-Akt

/Akt

PTEN-NCPTEN

Akt

p-Akt

GAPDH

miR-26a agomir + PTENmiR-26a agomir + PTEN-NC

PTEN-NC

PTEN

miR

-26a

ago

mir

+ P

TEN-NC

miR

-26a

ago

mir

+ PTEN

0.0

0.5

1.0

1.5

2.0

2.5

*

#

F

*

Rel

ativ

e lu

cife

rase

act

ivity

PTEN-3'UTR MUT PTEN-3'UTR WT0.0

0.5

1.0

1.5 NC mimicmiR-26a mimic

*

BA

Rel

ativ

e ex

pres

sion

of P

TE

N

Control brain tissue Glioma tissue0.0

0.5

1.0

1.5

*

C

Relative expression of miR-26a

Rel

ativ

e ex

pres

sion

of P

TE

N

0.03 4 5 6

0.2

0.4

0.6r = -0.525P = 0.001

D

Position 41-47 of PTEN 3’ UTR 5’ ...A C ACCAUGAAAAUAAACUUGAAU...

UCGGAUAGGACCUAAUGAACUU3’hsa-miR-26a-5p

Fig. 5 miR-26a activates the PI3K/Akt signaling pathway via down-regulation of PTEN. a the miR-26a binding site in PTEN 3’UTR; b detection ofluciferase activity; c PTEN expression in glioma tissues and control brain tissues; d correlation analysis of miR-26a expression with PTEN expression;e PTEN protein expression in HBMECs transfected with miR-26a agomir or miR-26a antagomir; f Akt protein level and phosphorylation level inHBMECs transfected with PTEN overexpression vector and/or miR-26a agomir measured using Western blot analysis; *, p < 0.05, compared withNC mimic (b), control brain tissues (c), agomir-NC (e), or PTEN-NC (f) group; #, p < 0.05, compared with antagomir-NC or miR-26a agomir + PTEN-NC group; All data were expressed as mean ± standard deviation; there were 46 cases of glioma tissues and 28 cases of control brain tissues; thecorrelation analysis was conducted using Pearson correlation analysis; comparisons between two groups were analyzed using unpaired t-test;comparisons among multiple groups were analyzed by the one-way ANOVA; the experiment was repeated three times; ANOVA, analysis ofvariance; HBMECs, human brain microvascular endothelial cells; miR-26a, microRNA-26a; NC, negative control; PTEN, phosphatase and tensinhomolog deleted on chromosome ten; Akt, protein kinase B; GAPDH, glyceraldehyde-3-phosphate dehydrogenase

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 11 of 15

Furthermore, our results also revealed that exosomalmiR-26a derived from GSCs elevated the levels of VEGF,MMP-2 and MMP-9, thus enhancing angiogenesis ofHBMECs. MMP-2 and MMP-9 are members of MMPfamily that is responsible for facilitating angiogenesis [56].As a major regulator of blood vessel formation and patho-logical processes, VEGF modulates proliferation, migra-tion and survival of ECs [57]. Due to its effect on EC

survival, VEGF elevation triggers glioma angiogenesis [58].Moreover, existing evidence suggests that miR-26a stimu-late tumor growth and angiogenesis in glioma [23], whichis consistent with our study that miR-26a in GSCs led tothe enhancement of tumor growth and angiogenesis invivo. In addition, tumor suppressor PTEN participates inregulation of tumor angiogenesis by negatively regulatingPI3K [59]. The activation of PI3K/Akt signaling pathway

Con

tent

of V

EG

F (

pg/m

l)

0

200

400

600

800

1000

*

A

Rel

ativ

e ex

pres

sion

of P

TE

N

0.0

0.5

1.0

1.5

*

B

Rel

ativ

e ex

pres

sion

of

p-A

kt/A

kt p

rote

in

GSC-EXs

miR

-26a

ago

mir

+

PTEN-NC

GSC-EXs

miR

-26a

ago

mir

+

PTEN

0.0

0.5

1.0

1.5

2.0Akt

p-Akt

GAPDH

C

Pos

itive

exp

ress

ion

of E

du r

ate

(%)

0

20

40

60 *

D

Tub

e le

ngth

(µm

)

12 h 24 h

12 h 24 h

12 h

24 h

Control

Control

GSC-EXsagomir-NC

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

GSC-EXsmiR-26a agomir

Control GSC-EXsagomir-NC GSC-EXsmiR-26a agomir

Control GSC-EXsagomir-NC GSC-EXsmiR-26a agomir

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

Control

GSC-EXsmiR-26a agomir + PTEN-NC

GSC-EXsmiR-26a agomir + PTEN

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

Contro

l

GSC-EXs

agom

ir-NC

GSC-EXs

miR

-26a

ago

mir

Control

GSC-EXsagomir-NC

GSC-EXsmiR-26a agomir

F

Num

ber

of b

ranc

hes

0

50

100

150

200

250

*

*

Rel

ativ

e pr

otei

n ex

pres

sion

VEGF

VEGF

MMP-2

MMP-2

MMP-9

MMP-9

GAPDH

0.0

0.5

1.0

1.5

2.0

2.5*

*

*

G

Cel

l num

ber

of m

igra

tion

0

50

100

150

*

E

25 µm 25 µm 25 µm

25 µm 25 µm 25 µm

25 µm 25 µm 25 µm

25 µm 25 µm 25 µm

*

0

500

1000

1500

2000

*

*

Conto

l

Fig. 6 Exosomal miR-26a promotes proliferation and angiogenesis of HBMECs through the inhibition of PTEN and activation of PI3K/Akt signaling pathway.a VEGF level in HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; b PTEN expression in HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; c Akt protein level and phosphorylation level in HBMECs co-cultured with GSC-EXsmiR-26a + PTEN or GSC-EXsmiR-26a + PTEN-NC; d proliferationof HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; emigration of HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; f tubeformation in HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; g VEGF protein level in HBMECs co-cultured with GSC-EXsagomir-NC or GSC-EXsmiR-26a agomir; *, p< 0.05, compared with the control or GSC-EXsagomir-NC group; all data were expressed as mean± standard deviation; comparisonsbetween two groups were analyzed using unpaired t-test; comparisons among multiple groups were analyzed by the one-way ANOVA; the experiment wasrepeated three times; ANOVA, analysis of variance; miR-26a, microRNA-26a; NC, negative control; PTEN, phosphatase and tensin homolog deleted onchromosome ten; Akt, protein kinase B; HBMECs, human brain microvascular endothelial cells; VEGF, vascular endothelial growth factor; GSC-EXsagomir-NC,exosomes derived from agomir-NC-transfected GSCs; GSC-EXsmiR-26a agomir, exosomes derived from miR-26a agomir-transfected GSCs

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 12 of 15

is also associated with the acceleration of metastasis andangiogenesis in glioma [1]. The in vitro experimentsconfirmed our findings suggesting that GSCs-derived exo-somes carrying miR-26a has the potential to enhanceangiogenesis of HBMECs by inhibiting PTEN and activat-ing the PI3K/Akt signaling pathway.

ConclusionIn conclusion, the aforementioned findings demonstratedthat exosomes secreted by GSCs transferred miR-26a into

HBMECs to promote the proliferation and angiogenesis ofHBMECs (Fig. 8). Therefore, exosomes from GSCs withup-regulated levels of miR-26a could be potential thera-peutic protocols for glioma. However, due to the fact thatthe research is still in the preclinical stage, the investiga-tion on the role and mechanism of exosomal miR-26a inthe angiogenesis and microenvironment of glioma in vivois insufficient. Therefore, more experiments are requiredto further explore the intrinsic mechanisms of exosomalmiR-26a in glioma.

Bra

nch

poin

ts

GSCagomir-NC

GSCagomir-NC

GSCmiR-26a agomir

GSCmiR-26a agomir

GSCagomir-NC GSCmiR-26a agomir

GSCantagomir-NC

GSCantagomir-NC

GSCmiR-26 aantagomir

GSCmiR-26 aantagomir

GSCantagomir-NC GSCmiR-26 aantagomir

0

50

100

150

*

#

CBA

Time (weeks)

Tum

or v

olum

e (m

m3 )

1 2 3 4 50

500

1000

1500

*

*

*

#

##

Con

tent

of V

EG

F (

pg/m

l)

0

100

200

300

400

500*

#

Num

ber

of C

D31

-MV

D

0

20

40

60

80

*

#

E FD

Rel

ativ

e pr

otei

n ex

pres

sion

PTEN

PTEN

VEGF

VEGF

GAPDH

0

1

2

3

GSCagomir-NC

GSCmiR-26a agomir

GSCantagomir-NC

GSCmiR-26 aantagomir

GSCagomir-NC

GSCmiR-26a agomir

GSCantagomir-NC

GSCmiR-26 aantagomir

GSCagomir-NC

GSCmiR-26a agomir

GSCantagomir-NC

GSCmiR-26 aantagomir

GSCag

omir-

NC

GSCm

iR-2

6a ag

omir

GSCan

tago

mir-

NC

GSCm

iR-2

6 aan

tago

mir

*

*

#

#

Tum

or w

eigh

t (g)

0.0

0.5

1.0

1.5

2.0

2.5

*

#

GSCagomir-NC

GSCmiR-26a agomir

GSCantagomir-NC

GSCmiR-26 aantagomir

GSCag

omir-

NC

GSCm

iR-2

6a ag

omir

GSCan

tago

mir-

NC

GSCm

iR-2

6 aan

tago

mir

GSCagomir-NC

GSCmiR-26a agomir

GSCantagomir-NC

GSCmiR-26 aantagomir25 µm 25 µm

25 µm 25 µm

Fig. 7 Overexpressing miR-26a in GSCs enhances angiogenesis in vivo. a the number of branches of blood vessels (5mm) around the CAM vehicle;b transplanted tumors in nude mice; c tumor volume and weight; d VEGF level in the serum of nude mice measured by ELISA; e protein levels of PTENand VEGF measured by Western blot analysis; f CD31-MVD detected by immunohistochemistry (× 400); *, p< 0.05, compared with the GSCagomir-NC

group; #, p< 0.05, compared with the GSCantagomir-NC group; all data were expressed as mean ± standard deviation; comparisons between two groupswere analyzed using unpaired t-test; comparisons among multiple groups were analyzed by the one-way ANOVA; the experiment was repeated threetimes; ANOVA, analysis of variance; miR-26a, microRNA-26a; NC, negative control; PTEN, phosphatase and tensin homolog deleted on chromosome ten;Akt, protein kinase B; VEGF, vascular endothelial growth factor; GSCs, glioma stem cells; CAM, chorioallantoic membrane; MVD, microvessel density; ELISA,enzyme-linked immunosorbent assay; GSCagomir-NC, agomir-NC-transfected GSCs; GSCantagomir-NC, antagomir-NC-transfected GSCs; GSCmiR-26a agomir,miR-26a agomir-transfected GSCs; GSCmiR-26a antagomir, miR-26a antagomir-transfected GSCs

Fig. 8 Internalization of exosomes secreted from GSCs carrying miR-26a promotes angiogenesis of HBMECs through the inhibition of PTEN and activatingthe PI3K/Akt signaling pathway via elevation of VEGF. PI3K, phosphatidylinositol-3 kinases; miR-26a, microRNA-26a; NC, negative control; GSCs, glioma stemcells; PTEN, phosphatase and tensin homolog deleted on chromosome ten; Akt, protein kinase B; HBMECs, human brain microvascular endothelial cells;VEGF, vascular endothelial growth factor

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 13 of 15

Additional file

Additional file 1: Figure S1. Identification of GSCs-derived exosomes.A, morphological features of GSCs-derived exosomes under the transmis-sion electron microscope; exosomes are round or oval membranous vesi-cles; B, particle size of exosomes determined using Nanosight; mostparticles have a dimension ranged from 30 to 150 nm; C, protein bandsof exosome surface markers (CD63, CD9 and CD81) detected by Westernblot analysis; D, expression of exosome surface markers (CD63 and CD81)detected using flow cytometry; GSCs, glioma stem cells. (EPS 7358 kb)

AbbreviationsANOVA: Analysis of variance; CAM: Chicken chorioallantoic membrane;CSCs: Cancer stem cells; DEGs: Differentially expressed genes; ELISA: Enzyme-linkedimmunosorbent assay; FBS: Fetal bovine serum; GEO: Gene Expression Omnibus;GFP: Green fluorescent protein; GO: Gene oncology; GSC: Glioma stem cell;GSCs: Glioma stem cells; HBMECs: Human brain microvascular endothelialcells; HRP: Horseradish peroxidase; KEGG: Kyoto Encyclopedia of Genes andGenomes; MES: Morpholine-ethanesulfonate; miRNAs: MicroRNAs; MMP: Matrixmetalloproteinase; Mut: Mutant; MVD: Microvessel density; MVECs: Microvascularendothelial cells; PBS: Phosphate buffer saline; PTEN: Phosphatase and tensinhomolog deleted on chromosome ten; PVDF: Polyvinylidene fluoride; RIPA: Radio-immunoprecipitation assay; SD: Standard deviation; SDS: Sodium dodecyl sulfate;Wt: Wild type

AcknowledgementsWe acknowledge and appreciate our colleagues for their valuable efforts andcomments on this paper.

FundingNone.

Availability of data and materialsThe datasets generated/analysed during the current study are available.

Authors’ contributionsFL designed the study. HW collated the data. Z-FW carried out data analysesand produced the initial draft of the manuscript. JD contributed to draftingthe manuscript. All authors have read and approved the final submittedmanuscript.

Ethics approval and consent to participateThis study was approved by the Ethics Committee of The Third XiangyaHospital of Central South University and all the included patients in the studysigned written consents. All animal experiments were conducted in strictaccordance with the Guidelines for Care and Use of Laboratory Animals.

Consent for publicationConsent for publication was obtained from the participants.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Received: 18 November 2018 Accepted: 17 April 2019

References1. Chang Y, Wu Q, Tian T, Li L, Guo X, Feng Z, Zhou J, Zhang L, Zhou S, Feng

G, et al. The influence of SRPK1 on glioma apoptosis, metastasis, andangiogenesis through the PI3K/Akt signaling pathway under normoxia.Tumour Biol. 2015;36(8):6083–93.

2. Liu M, Dai B, Kang SH, Ban K, Huang FJ, Lang FF, Aldape KD, Xie TX, PelloskiCE, Xie K, et al. FoxM1B is overexpressed in human glioblastomas andcritically regulates the tumorigenicity of glioma cells. Cancer Res. 2006;66(7):3593–602.

3. Ostrom QT, Bauchet L, Davis FG, Deltour I, Fisher JL, Langer CE,Pekmezci M, Schwartzbaum JA, Turner MC, Walsh KM, et al. Theepidemiology of glioma in adults: a “state of the science” review.Neuro-Oncology. 2014;16(7):896–913.

4. Li Q, Wang AY, Xu QG, Liu DY, Xu PX, Yu D. In-vitro inhibitory effect ofEGFL7-RNAi on endothelial angiogenesis in glioma. Int J Clin Exp Pathol.2015;8(10):12234–42.

5. Chan XH, Nama S, Gopal F, Rizk P, Ramasamy S, Sundaram G, Ow GS,Ivshina AV, Tanavde V, Haybaeck J, et al. Targeting glioma stem cells byfunctional inhibition of a prosurvival oncomiR-138 in malignant gliomas.Cell Rep. 2012;2(3):591–602.

6. Sun X, Ma X, Wang J, Zhao Y, Wang Y, Bihl JC, Chen Y, Jiang C. Glioma stemcells-derived exosomes promote the angiogenic ability of endothelial cellsthrough miR-21/VEGF signal. Oncotarget. 2017;8(22):36137–48.

7. Huang Z, Cheng L, Guryanova OA, Wu Q, Bao S. Cancer stem cells inglioblastoma--molecular signaling and therapeutic targeting. Protein Cell.2010;1(7):638–55.

8. Bao S, Wu Q, Sathornsumetee S, Hao Y, Li Z, Hjelmeland AB, Shi Q,McLendon RE, Bigner DD, Rich JN. Stem cell-like glioma cells promotetumor angiogenesis through vascular endothelial growth factor. Cancer Res.2006;66(16):7843–8.

9. Yin H, Shao Y, Chen X. The effects of CD147 on the cell proliferation, apoptosis,invasion, and angiogenesis in glioma. Neurol Sci. 2017;38(1):129–36.

10. Wang Y, Yan W, Lu X, Qian C, Zhang J, Li P, Shi L, Zhao P, Fu Z, Pu P, et al.Overexpression of osteopontin induces angiogenesis of endothelialprogenitor cells via the avbeta3/PI3K/AKT/eNOS/NO signaling pathway inglioma cells. Eur J Cell Biol. 2011;90(8):642–8.

11. Lang HL, Hu GW, Chen Y, Liu Y, Tu W, Lu YM, Wu L, Xu GH. Glioma cellspromote angiogenesis through the release of exosomes containing longnon-coding RNA POU3F3. Eur Rev Med Pharmacol Sci. 2017;21(5):959–72.

12. Lotvall J, Hill AF, Hochberg F, Buzas EI, Di Vizio D, Gardiner C, Gho YS,Kurochkin IV, Mathivanan S, Quesenberry P, et al. Minimal experimentalrequirements for definition of extracellular vesicles and their functions: aposition statement from the International Society for Extracellular Vesicles.J Extracell Vesicles. 2014;3:26913.

13. Kooijmans SA, Vader P, van Dommelen SM, van Solinge WW, Schiffelers RM.Exosome mimetics: a novel class of drug delivery systems. Int JNanomedicine. 2012;7:1525–41.

14. Johnsen KB, Gudbergsson JM, Skov MN, Pilgaard L, Moos T, Duroux M. Acomprehensive overview of exosomes as drug delivery vehicles -endogenous nanocarriers for targeted cancer therapy. Biochim BiophysActa. 2014;1846(1):75–87.

15. Arscott WT, Tandle AT, Zhao S, Shabason JE, Gordon IK, Schlaff CD, ZhangG, Tofilon PJ, Camphausen KA. Ionizing radiation and glioblastomaexosomes: implications in tumor biology and cell migration. Transl Oncol.2013;6(6):638–48.

16. Lasser C. Exosomal RNA as biomarkers and the therapeutic potential ofexosome vectors. Expert Opin Biol Ther. 2012;12(Suppl 1):S189–97.

17. Melo SA, Sugimoto H, O'Connell JT, Kato N, Villanueva A, Vidal A, Qiu L,Vitkin E, Perelman LT, Melo CA, et al. Cancer exosomes perform cell-independent microRNA biogenesis and promote tumorigenesis. Cancer Cell.2014;26(5):707–21.

18. Thind A, Wilson C. Exosomal miRNAs as cancer biomarkers and therapeutictargets. J Extracell Vesicles. 2016;5:31292.

19. Saadatpour L, Fadaee E, Fadaei S, Nassiri Mansour R, Mohammadi M, MousaviSM, Goodarzi M, Verdi J, Mirzaei H. Glioblastoma: exosome and microRNA asnovel diagnosis biomarkers. Cancer Gene Ther. 2016;23(12):415–8.

20. Otsuki T, Ishikawa M, Hori Y, Goto G, Sakamoto A. Volatile anesthetic sevofluraneameliorates endotoxin-induced acute lung injury via microRNA modulation inrats. Biomed Rep. 2015;3(3):408–12.

21. Ma ZL, Hou PP, Li YL, Wang DT, Yuan TW, Wei JL, Zhao BT, Lou JT, Zhao XT, Jin Y,et al. MicroRNA-34a inhibits the proliferation and promotes the apoptosis of non-small cell lung cancer H1299 cell line by targeting TGFbetaR2. Tumour Biol. 2015;36(4):2481–90.

22. Brower JV, Clark PA, Lyon W, Kuo JS. MicroRNAs in cancer: glioblastoma andglioblastoma cancer stem cells. Neurochem Int. 2014;77:68–77.

23. Qian X, Zhao P, Li W, Shi ZM, Wang L, Xu Q, Wang M, Liu N, Liu LZ, JiangBH. MicroRNA-26a promotes tumor growth and angiogenesis in glioma bydirectly targeting prohibitin. CNS Neurosci Ther. 2013;19(10):804–12.

24. Tuzesi A, Kling T, Wenger A, Lunavat TR, Jang SC, Rydenhag B, Lotvall J,Pollard SM, Danielsson A, Caren H. Pediatric brain tumor cells release

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 14 of 15

exosomes with a miRNA repertoire that differs from exosomes secreted bynormal cells. Oncotarget. 2017;8(52):90164–75.

25. Skog J, Wurdinger T, van Rijn S, Meijer DH, Gainche L, Sena-Esteves M, Curry WTJr, Carter BS, Krichevsky AM, Breakefield XO. Glioblastoma microvesicles transportRNA and proteins that promote tumour growth and provide diagnosticbiomarkers. Nat Cell Biol. 2008;10(12):1470–6.

26. Song MS, Salmena L, Pandolfi PP. The functions and regulation of the PTENtumour suppressor. Nat Rev Mol Cell Biol. 2012;13(5):283–96.

27. Georgescu MM. PTEN tumor suppressor network in PI3K-Akt pathwaycontrol. Genes Cancer. 2010;1(12):1170–7.

28. Smyth GK. Linear models and empirical bayes methods for assessingdifferential expression in microarray experiments. Stat Appl Genet Mol Biol.2004;3:Article3.

29. Chhabda S, Carney O, D'Arco F, Jacques TS, Mankad K. The 2016 WorldHealth Organization classification of tumours of the central nervous system:what the paediatric neuroradiologist needs to know. Quant Imaging MedSurg. 2016;6(5):486–9.

30. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB.Identification of a cancer stem cell in human brain tumors. Cancer Res.2003;63(18):5821–8.

31. Li Y, Cai B, Shen L, Dong Y, Lu Q, Sun S, Liu S, Ma S, Ma PX, Chen J. MiRNA-29bsuppresses tumor growth through simultaneously inhibiting angiogenesis andtumorigenesis by targeting Akt3. Cancer Lett. 2017;397:111–9.

32. Liu LZ, Li C, Chen Q, Jing Y, Carpenter R, Jiang Y, Kung HF, Lai L, Jiang BH.MiR-21 induced angiogenesis through AKT and ERK activation and HIF-1alpha expression. PLoS One. 2011;6(4):e19139.

33. Kim CJ, Terado T, Tambe Y, Mukaisho KI, Sugihara H, Kawauchi A, Inoue H.Anti-oncogenic activities of cyclin D1b siRNA on human bladder cancercells via induction of apoptosis and suppression of cancer cell stemnessand invasiveness. Int J Oncol. 2018;52(1):231–40.

34. Liu Q, Geng P, Shi L, Wang Q, Wang P. miR-29 promotes osteosarcoma cellproliferation and migration by targeting PTEN. Oncol Lett. 2019;17(1):883–90.

35. Zhang B, Liu Y, Zhang J. Silencing of miR-19a-3p enhances osteosarcomacells chemosensitivity by elevating the expression of tumor suppressorPTEN. Oncol Lett. 2019;17(1):414–21.

36. Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C,Dimeco F, Vescovi A. Isolation and characterization of tumorigenic, stem-likeneural precursors from human glioblastoma. Cancer Res. 2004;64(19):7011–21.

37. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW,Bigner DD, Rich JN. Glioma stem cells promote radioresistance bypreferential activation of the DNA damage response. Nature. 2006;444(7120):756–60.

38. Ostrom QT, de Blank PM, Kruchko C, Petersen CM, Liao P, Finlay JL, StearnsDS, Wolff JE, Wolinsky Y, Letterio JJ, et al. Alex's Lemonade StandFoundation Infant and Childhood Primary Brain and Central NervousSystem Tumors Diagnosed in the United States in 2007–2011. Neuro-Oncology. 2015;16(Suppl 10):x1–x36.

39. Magee JA, Piskounova E, Morrison SJ. Cancer stem cells: impact,heterogeneity, and uncertainty. Cancer Cell. 2012;21(3):283–96.

40. Hata K, Watanabe Y, Nakai H, Hata T, Hoshiai H. Expression of the vascularendothelial growth factor (VEGF) gene in epithelial ovarian cancer: anapproach to anti-VEGF therapy. Anticancer Res. 2011;31(2):731–7.

41. Chen Q, Jin M, Yang F, Zhu J, Xiao Q, Zhang L. Matrix metalloproteinases:inflammatory regulators of cell behaviors in vascular formation andremodeling. Mediat Inflamm. 2013;2013:928315.

42. Sampetrean O, Saya H. Characteristics of glioma stem cells. Brain TumorPathol. 2013;30(4):209–14.

43. Domenis R, Cesselli D, Toffoletto B, Bourkoula E, Caponnetto F, Manini I,Beltrami AP, Ius T, Skrap M, Di Loreto C, et al. Systemic T cellsimmunosuppression of glioma stem cell-derived exosomes is mediated byMonocytic myeloid-derived suppressor cells. PLoS One. 2017;12(1):e0169932.

44. Gong M, Yu B, Wang J, Wang Y, Liu M, Paul C, Millard RW, Xiao DS, Ashraf M,Xu M. Mesenchymal stem cells release exosomes that transfer miRNAs toendothelial cells and promote angiogenesis. Oncotarget. 2017;8(28):45200–12.

45. Li MY, Liu DW, Mao YG. Advances in the research of effects of exosomesderived from stem cells on wound repair. Zhonghua Shao Shang Za Zhi.2017;33(3):180–4.

46. Nilsson J, Skog J, Nordstrand A, Baranov V, Mincheva-Nilsson L, BreakefieldXO, Widmark A. Prostate cancer-derived urine exosomes: a novel approachto biomarkers for prostate cancer. Br J Cancer. 2009;100(10):1603–7.

47. Bobrie A, Colombo M, Raposo G, Thery C. Exosome secretion: molecularmechanisms and roles in immune responses. Traffic. 2011;12(12):1659–68.

48. Villarroya-Beltri C, Gutierrez-Vazquez C, Sanchez-Madrid F, Mittelbrunn M.Analysis of microRNA and protein transfer by exosomes during an immunesynapse. Methods Mol Biol. 2013;1024:41–51.

49. Figueroa J, Phillips LM, Shahar T, Hossain A, Gumin J, Kim H, Bean AJ, CalinGA, Fueyo J, Walters ET, et al. Exosomes from glioma-associatedmesenchymal stem cells increase the Tumorigenicity of glioma stem-likecells via transfer of miR-1587. Cancer Res. 2017;77(21):5808–19.

50. Li J, An G, Zhang M, Ma Q. Long non-coding RNA TUG1 acts as a miR-26a spongein human glioma cells. Biochem Biophys Res Commun. 2016;477(4):743–8.

51. Liu B, Wu X, Liu B, Wang C, Liu Y, Zhou Q, Xu K. MiR-26a enhancesmetastasis potential of lung cancer cells via AKT pathway by targetingPTEN. Biochim Biophys Acta. 2012;1822(11):1692–704.

52. Guo P, Nie Q, Lan J, Ge J, Qiu Y, Mao Q. C-Myc negatively controls thetumor suppressor PTEN by upregulating miR-26a in glioblastomamultiforme cells. Biochem Biophys Res Commun. 2013;441(1):186–90.

53. Dasari VR, Kaur K, Velpula KK, Gujrati M, Fassett D, Klopfenstein JD, Dinh DH,Rao JS. Upregulation of PTEN in glioma cells by cord blood mesenchymalstem cells inhibits migration via downregulation of the PI3K/Akt pathway.PLoS One. 2010;5(4):e10350.

54. Jiang L, Wang C, Lei F, Zhang L, Zhang X, Liu A, Wu G, Zhu J, Song L. miR-93 promotes cell proliferation in gliomas through activation of PI3K/Aktsignaling pathway. Oncotarget. 2015;6(10):8286–99.

55. Cai H, Xue Y, Li Z, Hu Y, Wang Z, Liu W, Li Z, Liu Y. Roundabout4 suppressesglioma-induced endothelial cell proliferation, migration and tube formationin vitro by inhibiting VEGR2-mediated PI3K/AKT and FAK signalingpathways. Cell Physiol Biochem. 2015;35(5):1689–705.

56. Ulrich D, Lichtenegger F, Unglaub F, Smeets R, Pallua N. Effect of chronicwound exudates and MMP-2/−9 inhibitor on angiogenesis in vitro. PlastReconstr Surg. 2005;116(2):539–45.

57. Chamorro-Jorganes A, Lee MY, Araldi E, Landskroner-Eiger S, Fernandez-Fuertes M, Sahraei M, Quiles Del Rey M, van Solingen C, Yu J, Fernandez-Hernando C, et al. VEGF-induced expression of miR-17-92 cluster inendothelial cells is mediated by ERK/ELK1 activation and regulatesangiogenesis. Circ Res. 2016;118(1):38–47.

58. Chen L, Li ZY, Xu SY, Zhang XJ, Zhang Y, Luo K, Li WP. Upregulation of miR-107 inhibits glioma angiogenesis and VEGF expression. Cell Mol Neurobiol.2016;36(1):113–20.

59. Bhattacharya D, Singh MK, Chaudhuri S, Acharya S, Basu AK, Chaudhuri S.T11TS impedes glioma angiogenesis by inhibiting VEGF signaling and pro-survival PI3K/Akt/eNOS pathway with concomitant upregulation of PTEN inbrain endothelial cells. J Neuro-Oncol. 2013;113(1):13–25.

Wang et al. Journal of Experimental & Clinical Cancer Research (2019) 38:201 Page 15 of 15